Improved hydrology for regional environmental prediction

In this report we document work aimed at improving the quality of the representation of land and rivers in multi-year, coupled atmosphere-land (UM-JULES) simulations over the British Isles. The approach taken was to use standalone (uncoupled) simulations of JULES to investigate the potential to improve the coupled system. The use of alternative soil ancillary information, generated by using a data assimilation framework and observations of soil moisture from the COSMOS-UK network to optimise the constants in a pedotransfer function, was found to result in improved simulations by JULES of river flow in a diverse sample of British rivers (as measured by standard statistics). The revised soil parameters tended to increase the variability of the simulated river on short timescales, and reduce variability on the annual timescale. In catchments with a large influence of slow baseflow the revised parameters tended to give poorer simulations, with too much variability on short timescales. A new representation of groundwater processes was implemented in JULES and applied, for the first time, across Great Britain. This was shown to allow an influence of groundwater on nearsurface hydrology and fluxes over large parts of the country. Modelled river flows were more realistic in many cases, though much of the improvement was due to differences in the representation of runoff generation rather than the introduction of groundwater. A more physically-complete parameterisation of river physics, using the local inertial equation, was applied in a range of catchments, again for the first time. Simulated river flows were improved in most cases. The final section of this report offers a perspective on how terrestrial hydrology and its impacts could be considered in the next generation of UK Climate Projections. The work reported here has significantly improved our capability in several areas of land surface modelling, to the extent that new developments could be tested in nationwide simulations of JULES. In each area the results are encouraging and further development will continue in the Hydro-JULES project